Μολὼν λαβέ

So the guys over at GBCS have finally  got around to responding,  in a piece titled 'On social class, anno 2014'.  Obviously the use of Latin makes the whole thing seem classy, but what the hell, let's  raise the cultural ante and go Greek!

 Like the Persians the Savage crew fight mob handed, but it is always good to see defections from the opposing ranks - three of the originals seem to have declined the opportunity to join in this round. I wonder why? Could it be that they are having second thoughts? I don't know but I think we should be told. [That's enough of the cliches. Ed.] 

But, lo, what is this? Replacing them are two new recruits. A hearty welcome to Laurison and Snee who seem to have invented the new sport of writing replies to responses to articles they didn't write themselves. Think about it, the mercenary possibilities are endless. I'm looking forward to a few guest gigs myself.

Seriously though, I will  in due course let y'all know what I think about their effort (you didn't seriously think I'd pass that opportunity over did you?) but not just yet. At the moment I'm rather busy working on things that don't involve pointing out the glaring deficiencies of logic and reason in the work of others. At least until the Easter break is over I'd like to preserve my positively cheerful mood and concentrate on construction rather than demolition.

I can give you a few teasers though. Firstly and trivially, I hope the version of anno 2014  sent to Sociology is better copy edited than the one on the web. For starters you really do need to sort out your references boys: McGovern, P., Hills, S., Mills, C., (2007) indeed. And while we are on the subject I'm still pondering the Freudian significance of the fact that though you cite my critique you don't actually manage to reference it. Whoa, that is seriously deep repression man.

And then there is the thing I'll pose as an Easter competition. You might want to have another look at the table of model selection statistics that appears immediately above Appendix 2. Let's call it Table 3 though you seem to have forgotten to give it a title. Can you spot anything odd about it?  Clue: why does classification error appear to increase as the number of estimated parameters increases? And why is the change not monotonic in the number of parameters? And while we are at it what does classification error mean in a model where some of the manifest indicators are continuous?


Jeez, you were supposed to be making things clearer, not digging an even bigger hole!